NOTES
Oct., 1957
1443
TABLE I EMPIRICAL CONSTANTS, MOLARREFRACTIONS, MOLARPOLARIZATIONS A N D DIPOLEMOMENTS OF DIBASIC ACIDSIN DIOXANE AT
Compound
e1
VI
a
25" b
Pa
MRD
Obsd.
' Ca1ad.a
2.63 2.37 14.5 156.5 0.97339 -0.423 Oxalic acid 2.2070 9.77 2.57 2.45 154.6 19.1 Malonic acid 2.2095 9.22 0.97320 -0.330 2.20 2.48 23.7 122.6 Succinic acid 2.2074 6.78 0.97337 -0.353 2.64 2.49 28.3 142.3 -0.260 Glutaric acid 2.2070 0.97323 7.65 2.60 2.49 33.0 171.7 Adipic acid 2.2082 7.83 0.97338 -0 283 3.08 2.88 Dithiodiglycolic acid 2.2105 13.87 0 97408 -0.677 233,9 39.8 Assuming free rotation about all C-C, C-S, S-S bonds but a ri id cis-planar configuration for the carboxyl group. However, free rotation about the S-S bond is unlikely in dithiodiglyc&c acid (see text). I
I
ing carbon atom, was chosen. These values of group moment and angle were based on the observed electric moment of propionic acid in dioxane,3 along with the known bond angles4 and bond moments' for the C-H (here taken as zero), C=O, C-0,and 0-H groups, slightly modified to take into account resonance in the carboxylic acids. Although the methyl group of methyl acetate lies 30" out of the O=C-0 plane4 it seems likely that the favored positions for the proton of a carboxylic acid will be in the O=C-0 plane hence the cisplanar configuration of the carboxyl group has been assumed in all the calculations. It may be noted that the bond moment and angle suggested' for the carboxyl group in benzoic acid (1.64 D. and 74') are nearly the same as those used in this article for the aliphatic carboxylic acids. The electric moments of the various acids were then computed assuming free rotation about all C-C, C-S and S-S bondss; these values are listed in Table I. The agreement with the observed values is fairly good, lending some support to the various assumptions made in the calculations (monomeric molecules in solution, cis-planar configuration of the carboxyl group and free rotation about C-C bonds). The observed value for the electric moment of succinic acid is surprisingly low and a similar minimum at succinic acid was observed in the series of diethyl esters.l The minimum observed at diethyl succinate was carefully confirmed by measurements in another solvent and in the vapor phase at various temperatures.l It is possible that this particular chain length favors a ring structure involving both carboxyl (or ester) groups of the molecule. Although the observed moment of dithiodiglycolic acid agrees reasonably well with the value computed for free rotation about all C-C, C-S and S-S bonds, it is most probablea that the C-S-S-C portion of the molecule is rigid with a dihedral angle about 70". If the C-S bond momenta is taken to be 1.27 D. and dihedral angles of 70, 80 and 90" are assumed for the C-S-S-C portion (along with free rotation about C-C and C-S bonds), the calculated values of the moment of the molecule are 3.10, 2.98 and 2.92 D., respectively. The observed momenf(3.08) thus agrees best with the value 70" for the dihedral angle, as found in di-n-butyl disulfide.6
Experimental The apparatus in this work was similar to that used by Hainsworth, Rowley and McInnesS to measure the effect of pressure on the e.m.f. of a cell containing a hydrogen electrode. The electrolysis cell consisted of two parts, of which the lower contained a normal calomel electrode, while the upper contained the test solution and the Pt-micro-
(3) L. G. Wewon, "Tables of Electric Dipole Moments," The Technology Press, Cambridge, Mass., 1948. (4) P. W. Allen and L. E. Sutton, Acta Cryst., 8, 46 (1850). ( 5 ) H.Eyring, Phys. Rev., 89, 746 (1932). (6) M. T. Roqers and T. W. Campbell, J . A m . Chsm, Boo., 74, 4742 (1952),
(1) Now Department of Chemical Engineering, University of Sydney. (2) H. A. Laitinen and I. M. Kolthoff, THIS JOURNAL,45, 1061 (1941). (3) W. R. Hainaworth, H. J. Rowley and D. A. McInnea, J . A m . Chem. SOC.,46, 1437 (1924).
Steric factors which apparently tend t o increase the dihedral angle in di-t-butyl disulfidea by about 10" are evidently not important here. Experimental Materials.-Dioxane was purified as described previously.? A sample of pure dithiodigl colic acid was kindly provided by Mr. Gavern T. Walker, Jondon, England. The remalning acids were commercial products which had been recrystallized twice and dried at 80". Oxalic acid was resublimed in vacuo a t 150' to obtain anhydrous material. Apparatus and Method.-The dielectric constants and densities of five solutions, ranging in concentration from about 0.0004 to 0.004 in mole fraction solute, were measured a t 25" by use of apparatus described elsewhere.? The constants el, a, Vl, b of the Halverstadt-Kumler equation are shown in Table I for each compound along wlth the computed molar polarizations P Zof the acids at infinite dilution.? The molar refraction M R D calculated from atomic and constitutive constants were used directly for the electronic plus atomic polarization term. The observed electric moments shown in Table I have a probable error about zt0.10.
Acknowledgment.-The author is indebted to Mr. Gavern T. Walker of London, England, for suggesting the study of dithiodiglycolic acid and for providing a sample of that compound. (7) M. T.Rogers, ibid., 77, 3681 (1955).
VOLTAMMETRY AT HIGH PRESSURE BY A. H. EWALD AND S. C. L I M ~ C.S.I.R.O., Division of Industrial Chemistry, High Pressure Laboratory, Unruewty o f Sydney8 Australia Received M a y 17, IQ67
Polarography using a dropping mercury electrode provides a useful tool both for the analysis of electrolyte solutions and for the investigation of electron transfer reactions, Laitinen and Kolthoff have shown that a platinum microelectrode can similarly be used to determine characteristic current-voltage curves. In the present note some experiments are reported in which such a Pt-microelectrode was used t o measure current-voltage curves at pressures up t o 3000 atm.
NOTES
1444
r
O.OIM IATM
10.0 ’
Vol. 61
TABLE I DIFFUSIONCURRENT A N D HALF-WAVE POTENTIAL OF Cu++Cuf REDUCTION I N A h . 0.5 N KCl MEASURED RELATIVE TO NORMAL CALOMEL ELECTRODE Pressure, cuso4, id, Ell2 atm.
1 1 7.5
*
2 a +-
5.0
8
rn
Pa.
0.01 ,005
11.1 5.5 9.0 7.5 4.5 6.7 5.3 3.8
1000 2000
.Ol .Ol
2000 3000 3000 3000
.005 .Ol .0075 .005
V.
0.120 .131 ,105 .077 .079
,059 .059 059 I
-
9
u
2.5
.
0.
+0.5
+0.25 0 -0.25 E.m.f. vs. n.c.e., v. Fig. 1. electrode. The calomel electrode was completely enclosed and electrical contact with it was made by a Pt-wire sealed through the cell wall; pressure was transmitted to it via the upper cell compartment. The Pt-microelectrode consisted of a short length of 30 gage platinum wire sealed into the end of a glass tube so that a/a2 in. of the wire was left protruding vertically into the solution. The two cell compartments were connected by a B10 standard taper joint; the male part of the joint was filled with a 3% agar jelly containing the supporting electrolyte. This acted as a salt bridge and prevented the contents of the two cell com artments from mixing. When pressure was applied to tge cell the jelly was extruded slightly from the joint and thus transmitted the pressure to the calomel cell. The test solution was sealed from the air and from the ressure transmitting oil by a layer of purified paraffin oil. ft was deoxygenated before the measurements by bubbling purified nitrogen through it. The whole cell was immersed in oil in a steel bomb which had two insulated electrodes and had previously been used for conductivity measurements.4 A Leeds and Northru potentiometer type K2 was used as the source of the variabg voltage and the current flowing through the cell was measured with a calibrated mirror galvanometer.
loglO(i/(id i)) against the applied e.m.f. The halfwave potential was found from these curves at log,,(i/(& - i)) = 0. The curves for the measurements at 1 atm. were straight lines with slopes of 16.2 volt-’, indicating that the reaction was reversible and involved the transfer of one electron.6 The plots for the measurements at higher pressures were curved at the ends, but all had a linear section of approximately the same slope at the center. The results show that pressure affects both the diffusion current and the half-wave potential of the reaction. The diffusion current of a Pt-microelectrode is given byz id
= nFACD/l
( 1)
where 7t is the number of electrons transferred in the reaction, P the Faraday, A the area of the electrode, C the concentration (mole l i t e r 1 ) and D the diffusion coefficient of the reacting ion, and 1 is the thickness of the diffusion layer, A refinement of this equation has recently been discussed by Pavlopou10s and Strickland.8 The decrease of id with pressure could be due to changes in D, C and 1. In an aqueous solution the volume concentration C will be increased by about 10% at 3000 atm,, while the effect of pressure on D is the reciprocal of the effect on the viscosity of the solution and corresponds to a decrease of approximately 23% at 3000 atm. The balance of the decrease in i d must then be due to an increase in the thickness of the diffusion layer. The change of diffusion current with concentration at constant pressure was found to be linear within the experimental error, as predicted by eq. 1. The results also show that the half-wave potential is independent of concentration, but is moved Results and Discussion 0.071 volt to a more negative value by an increase in The reduction of Cu++ to Cu+ in the presence of pressure of 3000 atm. This indicates that the re0.5 N KC1 was investigated. The solutions were actants are stabilized relative to the products and made up in conductivity water using A.R. quality that the reduction requires a greater driving force CuS04.5Hz0 which had been recrystallized, and at the higher pressure. Considering the various complexes formed by A.R. KC1. Measurements of diffusion currents could be reproduced within &lo% while the error copper in the presence of excess chloride ions, and the standard potentials involved in their reducin the half-wave potential was only =!=0.002volt. Some current-voltage curves are shown in Fig. 1, tions, it is thought that the observed wave correand Table I gives the diffusion currents i d and half- sponds to the reaction wave potentials El/, found a t various concentraCU++ f Hg + 3C1- +CuC12- + 1/2Hg&12 tions and pressures. The measured current was In this reaction ionic charges are neutralized and the corrected for the residual current, which flowed before the reduction wave started, and was plotted as effect of pressure is thus similar to that observed in (4) J. Buchansn and ( 1953).
El. D. Hamann, Trans. Faraday SOC.,4% 1425
( 5 ) I. M. Kolthoff and J. J. Lingme, “Polarography,” Interscience Publiiahers, Ino., New York, N. Y.. 1952, p. 213.
NOTES
Oct., 1957 many ionic systems,46$7 where it is generally found that an increase in pressure stabilize8 the state with the greater number of ionic charges. One can obtain a quantitative measure of this stabilizing effect from the approximate relation
aaQ/aP = -nFdE+/hP
1445
120 110 100
where AG is the free energy change of the reaction. Using the experimental values of El/,, we find bAG/ bP = 0.53 cal. mole-l atm.-I. Although the work described has been only of an exploratory nature, sufficient results have been obtained to show that voltammetry using a stationary Pt-microelectrode is feasible at high pressures and can yield useful information under these conditions. ( 6 ) 9. D. Hamann and W. Strauss, Trans. Faraday SOC.,61, 1684 (1955). ( 7 ) A. H. Ewald and 8. D. Hamann, Aust. J . Cham., 9, 54 (1956). (8) T. Pavlopoulos and J. D. Striokland, J . Electrochem. SOC.,104, 116 (1957).
THE SYSTEM 1,3$-TRINITROBENZENE2,4-DINITROTOLUENE BY LOHRA. BURKARDT Chemistry Division, U. S. Naval Ordnance Test Station, China Lake, California Received M a y 17, 1867
The system 1,3,5-trinitrobenzene-2,4-dinitrotoluene does not appear to have been previously investigated. This study of the system was made with an apparatus' which permitted a stepwise heating approach to the liquidus point. Solidliquid equilibrium a t each thermal step was assured by demonstrating constancy of the light transmission of the sample a t each thermal step. The 1,3,5-trinitrobenzene was recrystalliied from ethyl alcohol, washed with ethyl alcohol and air-dried. Before use, it was fused and allowed to crystallize under a vacuum twice. The melting point was then 123.6'. The 2 , 4 dinitrotoluene was recrystallized from hot ethylene dichloride after which it was fused and allowed to crystallize under a vacuum twice. It then melted at 80.2'. Six-gram samples of the required compositions were melted and stirred thoroughly. The temperature of the sample was allowed to fall until a small amount of solid was formed. The temperature was then raised step-wise. At each step the temperature was held constant until the light transmission of the sample became constant indicating equilibrium between the solid and liquid. In this manner the temperature was raised to a point at which a few crystals were in equilibrium with the liquid. The temperature was
90
5 j 80 70 60 50 0
10
20 30 40 50 60 70 80 1,3,5,-Trinitrobenzene, mole %. Fig. 1.
90 100
then raised in very small increments until by visual observation the last crystals disappeared. The temperature at this p o h t was taken as the liquidus temperature. The melting point of the eutectic was obtained by heating the completely solid sample through the eutectic melting point with a tem erature gradient of less than a tenth of a degree between t i e bath and Sam le. With such a small temperature gradient, a flat is ottained at the eutectic melting point, This system forms a eutectic mixture at 34 mole % of 1,3,5-trinitrobenzene. The eutectic melts at 51.3'. Data for this system is presented in Table I and shown graphically in Fig. 1.
ELECTRIC MOMENTS FROM EXTRAPOLATED MIXED SOLVENT DATA. 11. MOLECULAR ASSOCIATION1 BY GEORGEK. ESTOK AND S. P. SOOD Dept. of Chemistry and Chemical Enginesl.ing, Texas Technolooical Colleoe, Lubbock, Tezas Received May 2.8, lBb7
In a previous paper,2 which may be considered as Part I of this series, some preliminary results were reported on a method for obtaining electric moments of substances which cannot be studied directlv in benzene solution because of association or insufficient solubility. The moment in hypothetical benzene solution mag be calculated from data obtained by extrapolating the experimental values TABLE I MELTINQP O I N T DATAFOR THE SYSTEM 1,3,5-TRINITRO- obtained in mixed benzene-dioxane solvent, of varying composition, to pure benzene as solvent. BEN.EENE-2,4DlNITROTOLUJGNE It has been found necessary to modify the Eutectio 1,3,5-Trinitrobenzene. Me P mole % OC." m.p., OC. method previously reported, as will be indicated 0 70.2 later. Results are reported here on three compounds of decreasing association and benzene solu10 65.0 bility, respectively : benzoic acid, benzamide and 20 60.5 51.2 30 54.3 51.2 benzenesulfonamide. 40 50 60 70 80 90 100
60.0 73.9 86.7 97.4 106.3 114.5 123.6
(1) L. A. Burkardt, W. 9. McEwan and Inst., 2T, 693 (1956).
51.3 51.4
H. W. Pitman, Rev. Sci:
Experimental Preparation and Purification of Compounds.-Benzene (thiophene free) was refluxed overnight or longer with phosphorus entoxide, and then any low boiling material removed gy fractionation through a four foot, glass helix packed column. Constant boiling benzene was collected (1) This work was supported by a Frederiok Gardner Cottrell grant from Research Corporation, New York. N. Y. (2) G. K. Estok and C. H. Stembddge, J. Am. Chsm. Xoc., 76,4316
(1954).
